Chapter 3
Chapter 3
The Empirical Status of Binge Eating Disorder
Alexandra Dingemans
Patricia van Hanswijck de Jonge
Eric van Furth
This Chapter was previously published in: C. Norring, & R. L. Palmer (Eds.), 2005: EDNOS
eating disorders not otherwise specified. Scientific and clinical perspectives on the other eating disorders (pp. 6382). Hove, East Sussex: Routlegde.
Empirical status BED
Introduction
In the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders of
the American Psychiatric Association (DSM-IV) Binge Eating Disorder (BED) is proposed as
a new diagnostic category requiring further study and as an example of Eating Disorder Not
Otherwise Specified (ED-NOS)(American Psychiatric Association, 1994). The criteria are
described in an appendix, indicating that BED requires further research before it can be
incorporated as a fully accepted category in the DSM. No comparable diagnostic category
exists in ICD-10 (World Health Organization, 1992).
BED was introduced as a new eating disorder about ten years ago. The aim of this
chapter is to give an overview of the results and to discuss the empirical status of BED after
this decade of research. What is the empirical evidence for and against the state currently
defined as BED in the DSM-IV? In this chapter I will focus on the present status of BED as
an eating disorder. The issue of the relationship between BED and obesity will be discussed
elsewhere in the book.
The Diagnostic and Statistical Manual of Mental Disorders (DSM)
Classifying mental disorders by means of the DSM is one of the many methods of
classification. Since the publication of the third edition of the DSM the taxonomy proposed by
the American Psychiatric Association has become more dominant than anyone would have
believed possible in the light of the limited impact of the first and second editions (American
Psychiatric Association, 1952; American Psychiatric Association, 1968; Follette & Houts,
1996). In its proposal for the DSM-III the American Psychiatric Association considered mental
disorders as medical disorders (American Psychiatric Association, 1980). Although there is not
yet an agreement about the diagnosis and the criteria, in actual practice binge eating disorder is
already accepted as an eating disorder. However, there is considerable debate about how much
effort should be made to treat these patients in an eating disorder clinic. If patients with a
binge eating disorder are treated within an eating disorder clinic there is also discussion about
the kind of therapy, which is suited for these patients (Dingemans, Bruna, & van Furth, 2002).
Objections to this viewpoint came from the American Psychological Association. The
DSM reflects the underlying model of traditional medicine. In order to gain wide acceptance of
the system the task force of the DSM decided to abandon the theoretical (medical) view and to
cease referring to mental disorders as a subset of medical disorders. This decision largely
explains the syndrome-based and non-theoretical nature of the DSM (Follette et al., 1996).
According to the DSM-IV a mental disorder is defined as ‘a clinically significant
behavioural or psychological syndrome or pattern that occurs in an individual and that is associated with present
distress (e.g. a painful symptom) or disability (i.e. impairment in one or more important areas of functioning) or
with a significant increased risk of suffering death, pain, disability, or an important loss of freedom. In addition,
Chapter 3
this syndrome or pattern must not be merely an expectable and culturally sanctioned response to a particular
event, for example, the death of a loved one. Whatever its original cause, it must currently be considered a
manifestation of a behavioural, psychological, or biological dysfunction in the individual’ (page xxi). The
definition of dysfunction is crucial to the definition because something is not a disorder unless
something has gone amiss in the person concerned (Follette et al., 1996).
The proposal of Spitzer and others
In 1991 Spitzer and others suggested that BED should be included in the DSM-IV.
The rationale for their proposal was that many individuals with marked distress about binge
eating could not be diagnosed as having bulimia nervosa (BN). People with the BEDsyndrome have episodes of binge eating as do patients with bulimia nervosa, but unlike the
latter they do not engage in compensatory behaviours such as self-induced vomiting, the
misuse of laxatives, diuretics or diet pills, fasting and excessive exercise. The authors indicated
that such patients are common among the obese involved in weight control programmes
and/or belonging to overeaters anonymous (Spitzer et al., 1992; Spitzer et al., 1993). Although
the diagnosis BED was formulated with the obese in mind, obesity is not a criterion for BED.
For inclusion in a new version of the DSM, the ‘new’ diagnostic is required to
describe a pattern of symptoms not captured in the existing categories (Pincus, Frances, Davis,
First, & Widiger, 1992). Pincus and others (1992), who served on the DSM-IV Task Force,
stated that for a new category to be considered for inclusion in the DSM-IV “there must be
solid evidence that the diagnosis is useful in predicting prognosis, treatment selection or
outcome”.
In 1993 Fairburn and others considered whether Pincus’ arguments would apply to
the proposed addition of BED to the DSM. One argument against new diagnoses is that if
they are rare they may add unnecessary complexity to the already cumbersome system of
classification and be irrelevant for clinical use. Furthermore, incorporation of new categories is
likely to increase the overall prevalence of mental disorders. The addition of new and
unproven diagnoses carries the risk of trivializing the construct mental disorder and/or its
misuse. This is relevant for BED since we would not wish normal gluttony to be classed as a
psychiatric disorder (Fairburn et al., 1993b). The criteria proposed by Spitzer et al. (1992) have
been designed to minimize this risk.
However, Fairburn and others (1993b) argued that adding BED to the section of
eating disorders does not make it very complex because the present scheme is relatively simple.
A second argument against adding new diagnostic categories is that new diagnoses are
generally proposed by experts in the field concerned and are subsequently used by less expert
assessors who may identify more false positives. These inaccurate diagnoses may lead to faulty
treatments.
A third argument put forward by Fairburn and others is that ‘adding unproven
diagnostic categories may confer upon such categories an approval that they do not merit yet’.
Empirical status BED
They argue that there is no evidence to suggest that delineating BED from ED-NOS is a
useful or valid approach. This delineation “may impede efforts to devise better classificatory
schemes since investigators will inevitably tend to define their samples along the new lines”.
A fourth argument concerns the risk of definitional overlap across related categories.
Fairburn and others (1993b) argue that the delineation of BED could cause definitional
overlap between the non-purging type of bulimia nervosa and BED. Bulimia nervosa nonpurging type is distinguished from BED by the presence of compensatory behaviours (fasting
and excessive exercise) and/or undue emphasis on self-evaluation of body shape or body
weight. Although the existence of substantial differences between bulimia nervosa non-purging
type and binge eating disorder has been confirmed (Santonastaso et al., 1999), it is difficult to
draw clear boundaries between these two categories. It is unclear when these distinguishing
behaviours are severe enough to warrant the diagnosis of bulimia nervosa non-purging type
rather than binge eating disorder.
Another reason for adding a new diagnosis to the DSM, is to initiate research in the
field. The DSM-III and DSM-III-R have been facilitators of research in areas that would have
remained unresearched if they had not been included in an official nomenclature. Some see
this as one of the most important goals of the DSM. Others, however, think that research
should drive DSM and not the other way around. It makes no sense to include a category for
which there is no empirical support (Pincus et al., 1992). Once categories are included in the
DSM they are not easily deleted (Blashfield, Sprock, & Fuller, 1990). The insertion of BED in
the DSM-IV has led to a considerable increase in the amount of research into BED over the
last decade. The term binge eating disorder was first used in a paper that appeared in 1991. The
number of papers in which BED has been investigated has grown immensely over the past ten
years (see overview figure 1).
Figure 1: Overview of papers with ‘binge eating disorder’ as keyword
160
140
120
100
80
60
40
20
0
medline
PsychInfo
Web of Science
year
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
<1991
Embase
Chapter 3
Classification criteria
The Eating Disorders Work Group of the DSM-IV Task Force in conjunction with
Spitzer et al. (1992) developed preliminary criteria for the new eating disorder diagnosis
designed to identify “the many people who have problems with recurrent binge-eating, but
who do not engage in compensatory behaviours of bulimia nervosa, vomiting or the abuse of
laxatives” The criteria for the diagnosis of binge eating disorder were adapted from those for
BN , extra criteria being added to define the differences between the two disorders.
The definition of binge eating (criterion A) is identical for BN and BED. However, in
practice, binge eating among BED patients does not always conform to the requirement of the
consumption of “a large amount” of food during a “discrete period of time”, such as two
hours. Rossiter and others (1992) reported that many obese patients who overeat with a sense
of loss of control consume quantities of food that would not be described as ‘large’ during any
‘discrete period of time’ but would be considered excessive over the course of the day. This
pattern of overeating has been termed ‘grazing’ and is a frequent occurrence within the obese
binge eating population (Marcus et al., 1992).
Binge eating episodes are required to be associated with behavioural symptoms of
loss of control (criterion B). Spitzer and others (1992; 1993) felt that criterion B (symptoms of
loss of control) should be included so as to set a high threshold for the diagnosis of binge
eating disorder and to ensure that normal gluttony would not be classed as binge eating
disorder. However, this criterion does not seem to be fully distinct from criterion A (2) (a
sense of loss of control) or criterion C (feeling of distress regarding binge eating). Criteria B1,
B2, and B3 refer to aspects of loss of control whilst eating a large amount of food in a discrete
period of time (criterion A). Criteria B4 and B5 are related to characteristics of distress
described in criterion C.
The distress criterion was included in the DSM-IV proposed criteria for BED in
order to minimize false positives. Removal of this criterion would have increased the number
of individuals meeting criteria for the disorder by 10% in a weight control sample and by more
than 100% in a community sample in Spitzer’s field trail (1992). However, it is not clear what is
meant by ‘marked distress’. Distress may refer to emotional distress or impaired social or
occupational functioning as a consequence of binge eating behaviour. Does distress reflect the
patient’s self-report of his/her emotional state or does it also require social and/or
occupational impairment due to binge eating? Furthermore, de Zwaan (1997) stated that by
including the distress criterion one may identify individuals with high levels of general distress
which is not directly related to binge eating. For a diagnosis of BED, binge eating is required to
occur on two days a week (criterion D) whereas patients with BN are required to have two
episodes per week. This definition is based on the rationale that BED patients may have more
difficulty in recalling and labelling binge eating episodes than bulimia nervosa patients (Rossiter
et al., 1992). The end of a binge-eating episode in BN is often characterized by purging
behaviour, whereas in BED the termination of a binge episode is not punctuated by such
Empirical status BED
behaviour. However, defining the frequency of binge eating in terms of the number of days on
which the binge occurs seems to allow binges to be non-episodic, lasting as long as an entire
day.
The twice-weekly frequency criterion for binge eating episodes is arbitrary and has no
empirical support in cases of BN or BED. Various authors have found that raising the
frequency criteria of binge eating episodes from once or twice a week did not change the
pattern on measures of psychopathology or treatment outcome in binge eaters (Wilson, Nonas,
& Rosenblum, 1993; Striegel-Moore et al., 2000; Striegel-Moore et al., 1998; Garfinkel et al.,
1995).
Two binge eating days per week are required to occur during a six-month period in
BED whereas a three-month period is specified for BN The minimum duration of six months
is required in order to ensure that transitory episodes of binge eating are not diagnosed as
BED. Nevertheless, BED has been found to have a natural course that has a tendency to remit
(Hay, Fairburn, & Doll, 1996; Cachelin et al., 1999; Fairburn et al., 2000).
DSM-IV proposed diagnostic criterion E (which states that the diagnosis of binge
eating disorder should not be made if the patient engages in regular inappropriate
compensatory behaviours also seen in BN (self-induced vomiting, laxative abuse, fasting,
excessive exercise) or if he/she suffers from anorexia nervosa. However, there is no clear
definition of what is meant by ‘inappropriate’ compensatory behaviours. Some compensatory
behaviour in the obese is not necessarily inappropriate. Moreover, there is no definition for the
term regular. The term regular implies that some compensatory behaviour could be compatible
with the diagnosis of BED.
Unlike BN, the DSM-IV research diagnostic criteria for BED do not include ‘unduly
influenced self-evaluation by body weight and body shape’. It is not clear why this criterion has
been excluded from the proposed diagnostic criteria. However, Eldredge and Agras (Eldredge
& Agras, 1996) suggest that it may be due to uncertainty regarding the impact of the level of
obesity on such concerns. Although obesity is not among the diagnostic criteria for either
disorder, fewer BN patients are overweight than BED patients. The strong association
between BED and obesity may be related to a patient’s failure to compensate for the increased
calories consumed during the binge eating episodes. Despite omission of this criterion various
authors have observed a significant association between binge eating disorder and an
overconcern about body weight and body shape (Wilson et al., 1993; Cachelin et al., 1999;
Marcus et al., 1992; Masheb et al., 2000; Striegel-Moore et al., 2000; Wilfley, Schwartz, Spurrell,
& Fairburn, 2000b; van Hanswijck de Jonge, van Furth, Lacey, & Waller, 2003).
A comparison of characteristics of binge eating disorder and bulimia nervosa
Most studies that compared BED and BN used the criteria for BN in the DSM-III-R,
which made no distinction between the purging and non-purging subtypes of BN. The DSMIV does make such a distinction between these subtypes. Purging bulimics engage in self-
Chapter 3
induced vomiting, misuse of laxatives, diuretics or enemas. Non-purging bulimics do not purge
but do use other inappropriate compensatory behaviours (i.e. fasting or excessive exercise).
Several studies have compared BED patients with bulimics (with no distinction between
purging and non-purging subtype) (Marcus et al., 1992; Kirkley et al., 1992; Tobin et al., 1997;
LaChaussee et al., 1992; Goldfein et al., 1993; Mussell et al., 1995), and a few studies have
compared BED patients to purging (Fichter et al., 1993; Mitchell et al., 1999; Masheb et al.,
2000) and non-purging BN patients (Tobin et al., 1997; Hay et al., 1998; Santonastaso et al.,
1999). In order to interpret the results of these studies correctly it is important to know which
version of the DSM was used.
In a study in which normal weight subjects with BN (DSM-III-R) were compared to
obese subjects with BED, the latter seemed less anxious about their eating patterns and
bodyweight, felt less guilty about being overweight, were less preoccupied with their eating
behaviour, had a better overall opinion of themselves, were able to perceive internal states
more accurately, were more socially adjusted, and were more comfortable in maintaining
interpersonal relationships (Raymond et al., 1995). Both obese and non-obese BED subjects
have lower levels of dietary restraint than subjects with BN purging type (Masheb et al., 2000).
However, Marcus and others (1992) found that obese women seeking treatment for binge
eating reported levels of eating disorder psychopathology that were comparable to those of
normal weight BN patients (DSM-III-R). Similarly, a cross-sectional study (van Hanswijck de
Jonge, 2002) comparing BED to BN and obesity reported no significant difference between
the two eating disorders on overall restraint psychopathology, eating concern psychopathology,
body weight and body shape psychopathology. In all instances both BED and BN patients
scored significantly higher in eating disorder psychopathology than the non-binge eating obese
population.
Energy intake during an episode of binge eating seems to be different in BN (DSMIII-R) and BED. In a laboratory subjects were asked to binge on ice cream. Subjects with BN
(DSM-III-R) consumed four times as much as normal weight healthy controls (LaChaussee et
al., 1992). The same research group reported that subjects with BED ate only half the amount
of ice cream eaten by subjects with BN (Goldfein et al., 1993). One study compared the quality
and quantity of binges reported in individuals with BED and BN (Fitzgibbon et al., 2000).
Binges of subjects with BN included food that was higher in carbohydrates and sugar content
than the binges of subjects with BED. No difference was observed in the mean number of
calories consumed.
BED patients seem to show fewer comorbid psychiatric symptoms than BN patients
with either purging or non-purging subtype (Tobin et al., 1997). Schmidt and Telch (1998)
have documented higher levels of depression, impulsivity, self-defeating tendencies and lower
levels of self-esteem in bulimia nervosa than in binge eating disorder. Similarly, Raymond and
others (1995) reported higher levels of depression and anxiety in bulimia nervosa patients than
in binge eating disorder patients. Another study (Tobin et al., 1997) compared purging BN to
Empirical status BED
non-purging BN, BED and Eating Disorder Not Otherwise Specified (ED-NOS) on the
Hopkins Symptom checklist and a measure of borderline syndrome and depression. The BED
patients were reported to display significantly less anxiety, paranoia and psychoticism than the
other three groups. No other differences were found between the groups on general
psychopathology on the remaining measures. Unlike previous studies, no significant
differences were found between BED and the two BN subtypes. Another study (van
Hanswijck de Jonge, 2002) could not distinguish between BN and BED on general
psychopathology as measured by the SCL-90-R. Furthermore, no distinction was reported
between the two disorders on total levels of impulsivity. However, both BN and BED patients
scored significantly higher on both general psychopathology and levels of impulsivity than did
a group of non-binge eating obese patients. A study by Santonastaso and others (1999) showed
no difference between nonpurging bulimics and BED subjects on clinical and psychological
characteristics, such as psychiatric symptoms, frequency of bingeing, and impulsiveness traits.
However, on many of the variables, the BED group showed a significantly greater variance.
Webber (1994) does not document any significant differences between bulimia nervosa and
binge eating disorder.
Aetiology
In a community-based, retrospective case-control study, Fairburn and others
(Fairbu1998) aimed to identify specific risk factors for BED. They compared subjects with
BED with healthy controls, subjects with other psychiatric disorders and subjects with bulimia
nervosa. Their findings support the prediction that BED is associated with exposure to risk
factors that increase the risk of psychiatric disorder in general and that increase the risk of
obesity.
Little is known about the family characteristics of BED patients. One study found
that BED subjects rated their family environment as less supportive and cohesive, and less
engendering of direct and open expression of feelings than healthy controls. The BED group
scored worse than other eating disorder groups (Hodges et al., 1998). One study investigated
familial tendency for BED and the risk of other psychiatric disorders, but failed to show this
(Lee et al., 1999b).
In BN most individuals start dieting prior to the onset of binge eating (Mussell et al.,
1997; Marcus et al., 1995; Haiman et al., 1999). However, a fairly large subgroup of the
individuals with BED start binge eating prior to the onset of dieting (35-54%)(Grilo et al.,
2000; Abbott et al., 1998; Spurrell et al., 1997; Mussell et al., 1995). Dieting seems to play a role
in the etiology of BED, but research does not indicate that dieting is always a key factor in
BED, as it seems to do in BN (Howard et al., 1999). The binge-first group seem to diet
because they binge, not binge because they diet (Abbott et al., 1998). For subjects who start
binge eating before dieting, binge eating seems to be the primary symptom that leads to weight
Chapter 3
gain. Obesity is found to develop several years after the onset of binge eating (Haiman et al.,
1999; Mussell et al., 1995).
Course of binge eating disorder
Two studies have investigated the natural course of BED in the general population.
Fairburn and others (2000) followed 102 subjects with BED for five years. After five years
only 10% of these subjects still fulfilled the criteria for BED (1 subject (3%) fulfilled the
criteria for BN and 2 subjects (5%) for EDNOS . In total 18% had an eating disorder of
clinical severity. At the 5-year follow-up 77% of the group was abstinent (i.e. no objective
bulimic episodes). However, the group as a whole became heavier during the five years and a
large proportion tended to have a BMI over 30 (obesity) (22% at recruitment compared to
39% at follow-up). It was striking that only 8% had been treated for an eating disorder during
these five years.
Cachelin and others (1999) examined women with BED in the general population for
a period of six months. At the six-month follow-up 52% of these women suffered from fullsyndrome BED, whereas 48% appeared to be in partial remission. Treatment seeking in
general did not appear to be associated with improvement in BED over a relatively short time
period.
Fichter and others (1998) assessed the course and outcome of 68 women with BED
over a period of six years after intense inpatient treatment. In general, the majority of these
patients showed substantial improvement during treatment, a slight (in most cases nonsignificant) decline during the first 3 years after treatment ended and further improvement and
stabilization in the 4, 5 and 6 years following treatment. At the six-year follow-up only 6%
fulfilled the criteria for BED. In total 20% met the criteria for some eating disorder according
to the DSM-IV.
The studies mentioned above could be taken to indicate that treatment worsens the
course of BED, since a higher percentage of subjects improved without treatment. However,
subjects seeking help for BED seem to have more severe problems than subjects with BED in
the general population (Fairburn et al., 1996; Wilfley, Pike, Dohm, Striegel-Moore, & Fairburn,
2001). Furthermore, it is unclear from these studies whether objective binge eating behaviour is
replaced by overeating or subjective binge eating behaviour (explaining weight increase) in
those patients reporting abstinence from bingeing behaviour at follow-up.
Treatment
The treatment of BN has been researched extensively and there have also been
numerous controlled treatment studies (Schmidt, 1998); however, far less attention has been
paid to BED (Wilfley et al., 1997). Because BED is more similar to BN than to obesity without
binge eating, the first generation of BED treatment research focused on examining the efficacy
of those treatments that had been shown to be effective for BN: cognitive behavioural therapy
Empirical status BED
(CBT), interpersonal psychotherapy (IPT) and antidepressant medication. Many individuals
with BED however seek help for overweight. Treatment of obesity focuses on the reduction of
caloric intake, encourages a shift to a low fat diet, addresses any medical contribution to the
condition and initiates exercise. The underlying behavioural disturbances or the social and
psychological consequences of obesity are often neglected. If the treatment of obese subjects
with BED focuses only on the reduction of bodyweight and does not address binge eating or
underlying problems, binge eating continues or even worsens (Romano et al., 1995; Howard et
al., 1999). Weight-loss programmes seem to have little effect on the reduction of binge eating
in obese subjects with BED (Kirkley et al., 1992).
The studies that will be discussed here are those in which BED is classified by means
of the DSM-IV. To date there have been seven randomized controlled clinical trials conducted
in which the psychological treatment of BED has been evaluated (Agras et al., 1994a; Agras et
al., 1995; Eldredge et al., 1997; Peterson et al., 1998; Carter et al., 1998; Peterson et al., 2001;
Wilfley et al., 2002a).
Cognitive Behavioural Therapy (CBT) seems to cause a statistically significant
reduction in binge eating compared to no treatment (reduction in number of episodes after
treatment: 68-90%; abstinence from binge eating after treatment: 40-87%; reduction in number
of episodes after waiting list: 8-22%).
Two studies combined cognitive behavioural therapy (CBT) with weight loss
treatment (Agras et al., 1995; Eldredge et al., 1997). Treatment of obese subjects with BED
seems to be more successful if binge eating is treated before any attempts are made to lose
weight.
Two studies compared the efficacy of CBT versus Interpersonal Psychotherapy (IPT).
Agras and others (Agras et al., 1995) investigated the efficacy of IPT in treating overweight
patients with binge eating disorder who did not stop binge eating after 12 weeks of CBT.
Subjects who were successful after 12 weeks CBT received weight loss treatment. IPT did not
lead to further improvement in those who did not improve with CBT. Wilfley and others
(Wilfley et al., 2002a) randomized 162 overweight patients with BED to either CBT or IPT.
The frequency of binge eating dropped significantly in both groups after 20 weeks of treatment
(abstinence: CBT= 79% versus IPT = 73%) and at one-year follow-up (abstinence: CBT =
59% versus IPT = 62%). No differences were found between the two groups.
Nauta and others (2001) investigated the effectiveness of cognitive therapy and
behavioural therapy in a group of obese subjects with and without BED, who were recruited
from an obese community sample. Cognitive therapy appeared to be more effective than
behavioural therapy with regard to abstinence from binge eating at six months follow up (86%
and 44% respectively). At the end of treatment no differences were found in the abstinence
rates (67% and 44% respectively).
Two studies (Peterson et al., 1998; Carter et al., 1998; Peterson et al., 2001) examined
the efficacy of a self-help format (CBT) in the treatment of BED. Self-help formats seem to be
Chapter 3
effective (abstinence varied between 50% and 87%). However, some caution is needed with
the interpretation of the high abstinence rates, because the participants in these ‘self-help’
studies were probably less severely ill than those in the other studies (for example, subjects in
the study by Carter (1998) and others had not received any prior treatment).
Also a few double-blind placebo-controlled pharmacological trials have been
conducted in patients with BED. Drugs which have been examined are selective serotonine
reuptake inhibitors (SSRI´s; fluoxetine, fluvoxamine, sertraline)(Greeno et al., 1996; Arnold et
al., 2002; Hudson et al., 1998; McElroy et al., 2000), appetite suppressants (d-fenfluramine
(Stunkard et al., 1996) and sibutramine (Appolinario et al., 2003) and an anticonvulsant
(topiramate)(McElroy et al., 2003). Fluoxetine seemed to reduce dietary intake but did not
affect the frequency of binge episodes. This finding suggests that fluoxetine affects satiety, not
hunger (Greeno et al., 1996). In another study (Arnold et al., 2002) fluoxetine reduced binge
frequency significantly compared to placebo. Fluvoxamine was found to be effective in
reducing the frequency of binge episodes and in lowering Clinical Global Impression (CGI)
severity scores (Hudson et al., 1998). Sertraline seemed to be effective and well tolerated,
although the number of participants in that study was low (McElroy et al., 2000). DFenfluramine reduced the frequency of binge eating in obese women with BED, but failed to
reduce their bodyweight (Stunkard et al., 1996). A significant reduction of binge eating and
weight was found in the sibutramine group compared to the placebo group (Appolinario et al.,
2003). Topiramate was associated with significantly greater reductions in binge frequency
compared to placebo after 14 weeks (94% versus 46% repectively).
A striking finding in these pharmacological studies is a high placebo-effect. All studies
had a single-blind lead-in period from 1 to 4 weeks. After this lead-in period 42 to 44% of the
participants no longer met the DSM-IV-criteria for BED.
In all studies the drugs under investigation seemed to be more effective than placebo
with regard to the primary outcome measures. However, no long-term effects were found.
Further, drugs did not seem to bring about a reduction in bodyweight. Disadvantages of these
studies were the small number of participants and the short duration of the trials.
Cognitive behavioural psychotherapy is currently the most investigated treatment for
BED and consequently the treatment of choice for binge eating disorder.
Category or continuum?
The main question treated in this chapter is whether binge eating disorder can be
distinguished as a separate mental disorder. In order to define a distinct eating disorder, the
disorder must have well described characteristics. Is BED distinct from obesity? Another
important question is whether bulimic disorders are dimensional or categorical in nature?
In a few cross-sectional studies (Fichter et al., 1993; Kirkley et al., 1992; Howard et
al., 1999) patients with BED were compared to matched samples of patients with BN purging
Empirical status BED
type and of patients with obesity (BMI > 30). The scores of patients with BED had an
intermediate position between BN and obesity but were closer to BN than to obesity. In a
series of cross-sectional studies (van Hanswijck de Jonge, 2002) patients with binge eating
disorder were compared to patients with bulimia nervosa and non-binge eating obesity on
eating disorder psychopathology, general psychopathology, personality pathology (categorical
and dimensional). The study revealed a dichotomy between binge eaters (bulimia nervosa and
binge eating disorder) and non-binge eaters (obesity) rather than a continuum of severity
between the groups on all measures of psychopathology.
Williamson and others (1992) identified three relatively homogeneous subgroups of
subjects who had been diagnosed with EDNOS using two cluster analytic procedures. The
three atypical subgroups were contrasted with two groups of subjects with anorexia nervosa
and bulimia nervosa. These groups were very similar to the descriptions of sub threshold
anorexia nervosa, non-purging BN and BED. Subjects in the ‘BED’ group were morbidly
obese but did not report extreme motivation for thinness. They reported significant problems
with binge eating, including significant concern about loss of control over eating. These
subjects did not resort to extreme weight control methods such as purging or extremely
restrictive eating. Estimations of current body size were closer to norms and ideal body weight
preferences were larger than those predicted from norms. Members of the group also were less
biased in their assessment of actual and ideal weight than those of the other clinical groups.
They expressed a more realistic dissatisfaction with obesity.
In another study Williamson and others (2002) found further empirical support for
conceptualizing BN and BED as discrete syndromes. Three factors were found to account for
66% of the variance in eating disorder symptoms: binge eating, fear of fatness/compensatory
behaviours and drive for extreme thinness. The bulimia nervosa group scored high on the
features of binge eating and fear of fatness/compensatory behaviours but not on drive for
thinness. The binge eating disorder group scored high on binge eating but not on the other
two features. Furthermore, persons with a diagnosis of an eating disorder appeared to differ (at
least partly) from persons with no pathological eating behaviours in kind rather than simply in
degree.
Hay and others (1996) investigated the presence of clinically meaningful subgroups
among subjects with recurrent binge eating recruited from the general population. They
identified four subgroups by means of a cluster analysis. The results supported the concept of
bulimia nervosa and its division into purging and non-purging subtypes. The study failed to
provide evidence to support the construct “BED”. A possible explanation is that the
population under investigation was too young (16-35 years). Patients with BED seem to
present themselves for therapy in a later stage, when they are in their thirties and forties,
whereas subjects with the other eating disorders seek help when they are generally much
younger. The same population was re-analyzed by classifying the subjects according the DSMIV (Hay et al., 1998). A number of subjects were excluded from the analysis because they did
Chapter 3
not meet the DSM-IV criteria for any eating disorder. It was found that subjects with BN
purging type did not differ from those with BN non-purging type and the latter did not differ
from those with BED. There was a significant difference between subjects with BN purging
type and BED.
There seems to be some evidence that subjects who binge without purging are
different from subjects who binge and purge. Much research has focused on binge eating as
the core psychopathological feature of bulimic disorders. Some state that compensatory
behaviour ought to be the focal clinical feature of bulimia nervosa rather than binge eating
(Tobin et al., 1997). Few studies (O´Kearney, Gertler, Conti, & Duff, 1998; Garfinkel et al.,
1996; Walters et al., 1993; McCann, Rossiter, King, & Agras, 1991) have investigated the
differences between BN purging type and BN non-purging type. The overall evidence is that
there is a difference between these two subtypes. So far, no study has found evidence for the
division of bulimic eating disorders into the three distinct DSM categories such as BN purging
type, BN non-purging type and BED.
Others assume that bulimic disorders differ in degree rather than in kind. There is
some support for the notion that bulimic eating disorders exist on a continuum of clinical
severity, which starts with BN purging type (most severe), passes through BN non-purging
type (intermediate severity), and finishes with BED (least severe).
Discussion
The aim of this chapter has been to evaluate the empirical status of binge eating
disorder as defined by the DSM-IV. Ten years ago Spitzer and others (1991) reported that
there was no classification taxonomy in the DSM for the many individuals who engage in binge
eating but do not engage in inappropriate compensatory behaviours. The present discussion
considers Pincus’ arguments for and against the inclusion of a new diagnostic category in the
DSM following a decade of research in the field of binge eating disorder.
Firstly, Pincus argued that rare diagnostic categories might add unnecessary
complexity to the already cumbersome system of classification. However, epidemiological
studies have shown that 1 to 3% (Hay, 1998; Spitzer et al., 1993; Spitzer et al., 1992) of the
general population has binge eating episodes but does not engage in inappropriate behaviours.
The prevalence is higher in obese populations (1.3 - 70%). Furthermore, BED seems to be
more prevalent as the degree of obesity increases (Spitzer et al., 1992; Spitzer et al., 1993;
Ramacciotti et al., 2000; Basdevant et al., 1995; Ricca et al., 2000; Varnado et al., 1997).
Therefore, BED does not rank as a rare diagnostic disorder.
Secondly, Pincus argued that new diagnoses are generally proposed by experts and are
subsequently used by less expert assessors who may identify more false positives. Clinical
practice has indeed shown difficulties can arise in attempting to differentiate between binge
eating and emotional overeating in obese patients; such difficulties can lead to high false
positive diagnoses. A clear operationalisation of the criteria of a binge eating episode is needed.
Empirical status BED
Thirdly, Pincus argued that ‘adding unproven diagnostic categories may confer upon
such categories an approval that they do not merit yet’. Many studies have indicated that BED
does represent a distinct diagnostic entity. The characteristics of subjects with BED differ
significantly from those of subjects with bulimia nervosa and from those of obese subjects
without binge eating. Furthermore, taxonomic studies have shown that there is a distinct
category of BED, which differs from other clinical eating disorder categories. In day-to-day
clinical practice BED is a generally accepted category and various eating disorder clinics have
developed programmes for the treatment of BED.
Pincus’ fourth and last argument concerned the definitional overlap across related
categories. To date it has been difficult to distinguish BED from non-purging BN. Various
studies have shown that there is a significant difference between the BN purging subtype and
BED. Non-purging BN seems to occupy an intermediate position between these two
categories, not differing significantly from either of them (Hay et al., 1998). It is not yet clear
whether non-purging BN bears a closer resemblance to purging BN (supporting continued
classification under BN) or a closer resemblance to BED (supporting a merger of BED and
non-purging BN)(Striegel-Moore et al., 2000). Research is needed to clarify this issue.
Although there is evidence to suggest that BED represents a distinct eating disorder
category, the criteria as currently described would benefit from some major revision. Criterion
A as currently defined should be maintained. Binge eating episodes should be characterized by
the consumption of a large amount of food within a discrete period with a sense of lack of
control over eating. Criterion B seems to be superfluous. Criterion B measures binge eating
characteristics, which overlaps with criterion A (binge eating characteristics) and criterion C
(feelings of distress regarding binge eating). It is not clear what is meant by the term distress
(criterion C) as currently described. Distress may refer to an emotional state with regard to
binge eating or it may describe impairment in social or occupational functioning due to binge
eating. I suggest that both types of distress should be operationalised in the revised version of
criterion C. The DSM guidelines state that impairment in functioning is crucial in the definition
of any mental disorder.
For a diagnosis of binge eating disorder, binge eating is required to occur on two days
a week (criterion D) rather than in the form of two episodes per week. This is based on the
rationale that binge eating disorder patients may have more difficulties in recalling and labelling
binge eating episodes due to the absence of purging behaviours punctuating the termination of
an episode. However, counting the number of days allows binges to last an entire day (in
theory). In a population characterized by a high percentage of compulsive overeaters (without
loss of control), this may complicate the separation of the diagnoses of binge eating episodes
and compulsive overeating. Therefore, I suggest counting the number of binge eating episodes
rather than counting the number of days.
Chapter 3
I suggest eliminating the term ‘regular’ as mentioned in criterion E. The elimination
of this term ensures clearer boundaries between bulimia nervosa and binge eating disorder.
Existing studies support the need for cognitive criteria in addition to the existing
behavioural diagnostic criteria for binge eating disorder. Hitherto, various existing studies have
argued for the inclusion of overconcern with body weight and body shape in self-evaluation
(Eldredge et al., 1996; Wilson et al., 1993; Wilfley et al., 2000b).
Furthermore, future research needs to clarify the impact of obesity on the
psychopathology of BED and vice versa. Although obesity is not a criterion for the diagnosis
of BED, the classification for BED was created with the obese in mind (Spitzer et al., 1991).
Future research will show whether obesity should be admitted as a criterion for BED in the
same way as underweight was included as a criterion for anorexia nervosa.
In summary, I believe that BED represents a distinct eating disorder category and
suggest that it be admitted into the next version of the DSM.
Empirical status BED
Chapter 3